Cloud Hosts:
We've discussed how small businesses can benefit from cloud hosting before, with many small- to medium-sized businesses putting to work the hosting method that provides them with exactly what they need for their website. Now, it seems large corporations are catching on to the cloud's usefulness for their own websites. 
Analysts are now predicting all areas of cloud hosting services, whether platform-, software-, or infrastructure-as-a-service, will grow to an estimated $60 billion in the next three years. As such, cloud hosting providers are beefing up their investments into not only hosting services, but colocation and managed unified communications as well.
How Cloud Hosts Stand Out Among The Rest
With so many cloud hosting providers out there, what can one company do to stand out above the rest? It's simple: better network monitoring.
A recent survey of IT organizations conducted by Internap, a cloud hosting and colocation services provider, points out that 88% of businesses and other organizations gain network usage and application performance trend data. Other studies indicate that these companies will pay top dollar for these services, such as application optimization, application transaction response time performance monitoring, and end-to-end correlation.
Network Performance Is Key
When switching to cloud hosting, numerous issues regarding application performance and other problems can arise. Latency, jitter, and packet loss found in cloud hosting become quite evident. Take virtual desktop infrastructure (VDI) as an example: these applications rely on transactions being conducted, as well as messages being sent, in an instant. Intermittent packet loss or latency issues in the network at any time can be costly in terms of productivity for any business.
Network monitoring is the solution. But, do most cloud hosting organizations have the time or money for that? Probably not. Also, take into consideration that most traffic travels between remote offices rather than hitting the data center monitoring point. You can see that cloud providers would like a little help in this area!
What Can Be Done
It is obviously imperative we gather information regarding the networks that keep the cloud up and running. Customers want to get exactly what they're paying for, and hosting providers want to be able to respond in an instant to outages and issues with application performance.
So what can a cloud provider do? Here are some problems network operators can come across in attempts to get a decent cloud platform up and running smoothly, with ideas as to what they can do about it:
- Determining both the location and cause of poor application performance. Figure out if it's the application, or the network. If the network is to blame, figure out what's causing it (server, storage, or WAN).
- Large amounts of traffic affecting site performance in numerous areas. Packet capture and analysis can sometimes work in data centers or colocation, but if you want the branch office of your IT organization to see what's going on, you'll need to figure out a new way to gather data regarding protocols, applications, and networks without relying on additional equipment.
- Planning for disaster before the application is even implemented. If you want a smooth experience when adding an application, you need to conduct predictive monitoring to assure the current network infrastructure can handle the change you are going to make. An example: before you move a virtual machine or server from data center to data center, look at how that might impact service, traffic, and overall user experience.
The bottom-line: a cloud can only be effective if the network is effective. If you're having issues with the cloud hosting services you're offering, start at the network and work your way up from there. Analyze its performance, and if improvements need to be made, you'll likely see a big difference in the quality of services you're offering.
Are you a cloud hosting provider looking to improve upon the services you offer? Have you thought to look to the network first, or did this seem to be common sense?
